[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
## Re: [igraph] the old "cannot reserve space for vector" problem

**From**: |
Pascal Jürgens |

**Subject**: |
Re: [igraph] the old "cannot reserve space for vector" problem |

**Date**: |
Tue, 8 Jun 2010 10:12:47 +0200 |

Hello,
sorry to hijack this thread, but my question is related.
Are there any practical limits to using shortest_paths() with much larger
networks? Especially, are there memory leaks and / or other limits when using
the python interface?
I'd like to compute it on a network with 80k vertices. Since the algorithm has
quadratic memory requirements, this results in around 50GB of RAM. As long as
there are no serious leaks, this is a small enough amount to run on a quad
memory EC2 instance. Using the unweighted algorithm (is it Johnson's?) with
O(n(V+E)), this seems feasible. Or will the time complexity bite me before the
RAM does?
Pascal
On Jun 5, 2010, at 11:29 , Tamas Nepusz wrote:
>* By the way, the shortest path matrix of a graph with 18209 vertices has *
>* 331567681 elements; if each element takes up 8 bytes, then the total memory *
>* requirement for storing the whole matrix at once will be about 2.47Gb, so the *
>* trick suggested by Łukasz might also work.*